The efficiency of gene targeting is limited, at least in part, by the inaccessibility of the chromosomal target to cellular recombination machinery. Studies in a variety of organisms have demonstrated that intentional cleavage of the target greatly stimulates the desired homologous recombination event. At present there are no reagents that can efficiently introduce a double-strand break at an arbitrarily chosen target site. The goal of the proposed study is to test the capabilities of a novel class of designed nucleases that have great promise as initiators of targeted recombination. These enzymes are engineered hybrids between the nonspecific DNA cleavage domain of the FokI restriction endonuclease and DNA recognition domains of the Cys2His2 zinc finger family. These chimeric nucleases have been shown to direct cleavage in vitro to sites recognized by the zinc fingers. Because zinc finger DNA-binding domains can be designed or selected to recognize a wide variety of DNA sequences, these chimeras should allow cleavage to be directed to many different targets. The resulting procedures will have immediate applicability to directed genetic manipulations for gene therapy and for experimental investigations of gene function. Initial aspects of the proposed study will focus on further characterization of target recognition and discrimination by the chimeric nucleases in vitro. For comparison with in vivo experiments, it is important to know how the enzymes operate in the absence of chromatin proteins, and optimum recognition specificity will be crucial to ultimate genomic cleavage. Sites for the chimeric nucleases will be placed in DNA substrates that can test homologous recombination capabilities in Xenopus oocytes. These DNAs will be introduced into oocytes, where they are assembled into chromatin, followed by injection of the corresponding nuclease. Both cleavage and recombination of the DNAs will be evaluated. The advantages of Xenopus oocytes for these studies derive from their large capacity for recombination, ease of substrate manipulation, and the possibility of characterizing reaction intermediates. Once the basic properties of the chimeric nucleases have been determined, new versions with zinc fingers targeted to novel sites will be generated by randomization and selection techniques. This will prepare the way for targeting of specific sites in complex genomes.